Mains frequency harmonics testing

S l i d e 1Tuesday, 01 February 2000Harmonic Current EmissionsPhilip W. Carter a CE mark europe ltdA simple low frequency test or a complex nightmare? An overview of the current status of the EMC StandardPart 3: Limits – Section 2:Limits for harmonic current emissions(Equipment input current £16 A per phase)S l i d e 2Tuesday, 01 February 2000Introduction•IEC/EN 61000-3-2 introduced in 1995–Main issues with Class D products•Number of Ambiguities–Confusion over Class D wave -shape»Envelope shape »Each Half Period–Confusion over basis for Class D limits»Complex interpretation »Rated Load Condition »Measured LoadThe International Standard was published in 1995 and has since developed into the most controversial of all the EMC standards.In the main the problems are mostlyrestricted to Class D products. A number of areas are poorly defined within the document and have lead to a variety of interpretations. S l i d e 3Tuesday, 01 February 2000Introduction• A very ambiguous standard–Several apparently valid interpretations•Most test systems now allow various interpretations to be set up as defaults–No Guidance available for which to choose –Need for consistent approachThe muddle was clear from the first attempts to accredit laboratories to this standard where it became evident that there were several very differentimplementations and many shortcomings to conform to the requirements.Although many of the test systems have evolved significantly since they were introduced and can now demonstrate various options, they can still havesignificant differences in approach, method and results when testing transitory harmonics.S l i d e 4conditions given in the relevant clause of annex C.Some of the ambiguities will become clear if we take a look at the current standard. As you can see there are four Classes of equipment. The last Class, ‘D’ has a definition that requires the wave shape to meet the definition of figure 1 the active input power to be measured and for it to be less than 600Watts. S l i d e 5Overview of standardEquipment shall be deemed to be Class D if, under the test conditions given in annex C, the input current wave shape of referred to its is within the envelope as given in figure 1 duration of each half-period ; this implies that wave forms having small peaks outside the envelope are considered to fall within the Envelope. The centre line, M, coincides with the peak value of the input current.When we read each statement within the standard we must decide a. What it means and b. whether we comply with it. From an Assessor viewpoint the onus is me to ask these questions from the point of Black and White. This comes from the point of view that if it is not wrong it is right.So from this what does each half cycle mean? S l i d e 6Overview of standard6.2 Harmonic current measurementWhen limits for harmonic currents are given as a function of the fundamental current or active input power, input power shall be measured under the same conditions The limits are applicable to steady-state harmonic currents measured according to annexes A, B and C.Here we have another statement that is possibly confusing. “The currents and input power shall be measured under the same condition.” Some exponents of the complex solution argue this can only mean a simultaneous measurement of current and voltage throughout each 16 cycle (320mS) window and calculate an active input power from this. Then apply this active input power to calculate a limit. This limit is then applied to the Harmonic currents measured in the same window.Where does it say that?S l i d e 7Tuesday, 01 February 2000Overview of standard–6.2.2 Transitory state•For transitory harmonic currents, measured according to annexes A, B and C, the following applies:–a) harmonic currents lasting for not more than 10 s when a piece of equipment is brought into operation or is taken out of operation, manually or automatically, are disregarded;This is another area where most original implementations were unable to comply with the standard requirement to ignore up to but no more than 10 seconds.S l i d e 8Seems innocuous enough but the proper implementation of this filter has not been demonstrated for all analyser solutions.S l i d e 9Overview of standard7 Harmonic current limitsThe following limits apply, with the provision that the limits f high-power equipment (>1 kW) for professional use are kept under 7.1 Limits for Class A equipmentFor Class A equipment, the harmonics of the input current shall not absolute values given in table 1.7.2 Limits for Class B equipmentFor Class B equipment, the harmonics of the input current shall not maximum permissible 7.3 Limits for Class C equipmentQ.What is the difference between ‘absolute values’ and ‘maximum permissible values’? A.It seems they mean the same thing but it keeps us guessing!S l i d e 10Here come the next two questions:What does ‘rated load condition’ mean?When is the implementation date of this standard?(Given that it becomes mandatory in January 2001 does that mean 2005?) S l i d e 11Tuesday, 01 February 2000Class D LimitsThe table giving the limits applicable to Class D equipment shows how the limits are proportional to the power and given in terms mAmps of Harmonic current per Watt.Note that Class D equipments do not have any even Harmonic limits. S l i d e 12Tuesday, 01 February 2000Overview of standard• B.1 General•Any type of wave analyser may be used, for example,frequency -domain instrumentation using selective amplifier, heterodyne, multiple passive filters, spectrum analyser tuned to the frequency to be measured and time-domain instrumentation using digital filters or Discrete -Fourier-Transform (DFT). The instrument may be of either the indicating or the recording type. The following requirements ensure the equivalence offrequency -domain and time-domain instruments in a practical sense.The Standard clearly and unambiguously maintains that the analysers may befrequency or time domain. Specifically it may be a spectrum analyser tuned to the harmonic frequency of interest. And also defined that it may be an indicating or recording type.S l i d e 13Another significant point coming from Annex B is that there is no preferredmeasurement method and more importantly no reference measurement method.S l i d e 14Annex CThe test conditions for the measurement of harmonic currents associated with some types of equipment are given in theFor equipment not mentioned there, users’ operation controls or automatic programmes shall be set to produce the maximum harmonic components under normal operating conditions for each successive harmonic component in turn.The equipment is tested as presented by the manufacturer.Preliminary operation of motor drives by the manufacturer may be needed before the tests are undertaken to ensure that resultscorrespond with normal use.Annex C gives the definitions for the measurement configuration for various types of equipment.This particular clause however makes a requirement that if the equipment is not included in the Annex then a series of measurements need to be performed to ascertain the maximum harmonic currents for each harmonic in turn all (40 of them). This is an incredibly complex requirement and could take a test laboratory many days of investigation to identify the true maximum harmonic currents.When we wanted to change this clause we were told by a previous convenor that it was the only way to get the manufacturers to the negotiating table.No wonder compromise is not coming easy!S l i d e 15Tuesday, 01 February 2000Annex C10 ITE–C.10 Test conditions for information technology equipment (ITE)•ITE is tested with the equipment configured to its rated current . In this case, the equipment, if necessary, may be configured with its power supplies loaded with additional load (resistive) board s to simulate rated current conditions.•For ITE systems designed for use with a manufacturer-supplied power distribution system, e.g. transformers, UPS, power conditioner, etc., compliance with the limits of this standard shall be met at the input to the power distribution system.For ITE equipment the annex specifies that the equipment be tested configured to its rated current.For some equipment this would mean the manufacturers would need to makehundreds of different cards to simulate the variety of configurations possible.S l i d e 16Tuesday, 01 February 2000EMCTLA Experiment•Attempt to quantify problem–Do the various equipments used by the test laboratories give different measured results?–Do the various equipments used by the test laboratories give different limits?–Do the test laboratories make different interpretations of the standard?Clearly there are a number of issues to concern the testing laboratory as well as the manufacturer. It was decided within the EMC Test Laboratories Association in the UK to conduct a laboratory inter-comparison exercise to see how results would compare between the installed base of test solutions.The three basic questions were addressed in the comparison by limiting the set-upprocedure supplied to the equipment under test only.S l i d e 17Three basic equipment under test conditions were evaluated the firstcompared the steady state results from the different laboratories.The results shown indicate an extremely good correlation between the different test analysers. S l i d e 18EMCTLA ExperimentTest 1 Limits791Harmonic NumberWhen we review the limits chosen by the laboratories to apply for this test we do see three limit values that do not fall into line with the expected pattern.Generally however the correlation crossed the boundaries of test systems and the three odd results were not strictly associated with a particular test system type.S l i d e 19This is a test where the programmable load is modulated with a pulse waveform to give two distinct states or levels.This represents a fluctuating load with levels of 2 to 4 Amps, 0.25 Hz, 10% duty cycle. The transient increase in current to 4 amps occurs once every 4 seconds and lasts for 400mSecondsAgain we have a fairly good correlation of results. S l i d e 20EMCTLA ExperimentTest 2 Limits57910Harmonic NumberHere we can see the limits chosen by the test systems for these results to be compared against.There are two fairly distinct levels chosen for limits, approximately 300mA and approximately 500mA.Some of the histogram bars are blank where the results fell between the 100% and 150% amplitude levels for more than 10% of the measurement time.The laboratories concerned could not provide these limits. They may not be available due to the need to summarise such a vast amount of data.S l i d e 21EMCTLA ExperimentTest 3 Results7910Harmonic NumberThe results for the third test a moresignificant fluctuation fluctuating load of 2 to 7.5 Amps, 0.25 Hz, 20% duty cycle this time not only a higher transient current is simulated but also an 800mS transient time.Still the measured results are surprisingly consistent.S l i d e 22Here we can really see a differencebetween the various approaches taken by the test system suppliers. The limits that have been chosen are between 400mA and 850mA for the 3rd Harmonic current level.Some of the histogram bars are blank again where the results fell between the 100% and 150% amplitude levels for more than 10% of the measurement time. S l i d e 23Tuesday, 01 February 2000EMCTLA Experiment•Preliminary Conclusions–Do the various equipments used by the test laboratories give different measured results?•Very Good Correlation for Steady State, and still reasonable for transitory harmonic tests!–Do the various equipments used by the test laboratories give different limits?•Significantly over 100% variation, many limits not known by laboratory!–Do the test laboratories make different interpretations of the standard?•Tend to rely on the defaults of the software!The results of the initial analysis indicate that there are indeed problems in testing to the current test standard. Although we have a surprisingly good correlation between measurement solutions for the actual harmonic currents we have significant differences in the calculated limits to apply to the results achieved.These results have been achieved with little intervention in the test system setting from those that represent the default values. A great reliance has been made on the test system manufacturers having correctly interpreted the standard.Understandably in this case we see the ambiguity showing through.S l i d e 24Tuesday, 01 February 2000EMCTLA Experiment•Areas of controversy not explored by experiment.–Wave-shape interpretations•Each half cycle •Half cycle areaThe current experiment has not evaluated differences due to the other areas of controversy. It is only speculation how much these would have an effect, but it must be considered likely that from the defined envelopes in some test solution providers that some equipment would not consistently be allocated into class D.Some of the approaches considered clearly implement the complex analysis of limits. Frequency domain solutions have not yet been analysed, however it is clear that such techniques would not agree on limit values established by the complex methods.S l i d e 25Tuesday, 01 February 2000Wave-shape InterpretationsThere are two specific issues associated with the analysis of the wave-shape and comparing the shape measured with the envelope.If the wave shape goes out of the envelope for any half cycle during the measurement period, does that put the equipment into Class A?If we have any DC offset on the waveform and the current waveform falls below the baseline shown in the envelope for more than 5% does this then fall into Class A? Intended or not this is the only black and white interpretation of this standard.S l i d e 26 Tuesday, 01 February 2000Complex Vs Simple•Complex approach–Different limit for each 320mS (proposed 200mS) time window equating to up to 469 limits per harmonic during the test (proposed 750). Giving 18,760 separate test limits for each test (proposed 30,000).–For fluctuating harmonics double the above numbers or add another 1876 limits (proposed 3000) to the final figure when 10% values are calculated.•Simple approach–Use Rated Power if available–Measure maximum current and calculate 1 limit for each HarmonicWhat justification can there be for making the measurement of power line Harmonic currents the most complicated analysis in EMC history? 18, 760 separate limits to compare our results to.It also begs the question of how many of these 18760 different limits would be the same when measured at another laboratory. What would it mean to a manufacturer to be told that he fails in the 353rd time window by 7mAmps, when the time window before was a higher value but an even higher limit so it passed. How would he go about investigating that?Surely even common sense requires a complete rethink and getting back to some simple limits.S l i d e 27Tuesday, 01 February 2000Frequency Domain Measurement•Would it be possible to measure the fluctuating limits in each 320mS window with a Spectrum Analyser–Could we establish 18,760 limits?–Could we see the 10% time period allowed higher fluctuations and establish further 18,760 *150% limits•Could they be equivalent methods–with the complex interpretation?I do not propose to even try to suggest a method of using a frequency domain instrument such as a spectrum analyser to try and emulate the complex approach.The standard clearly mandates that the methods are equivalent.Have the committees inadvertently or otherwise mislead some test equipment manufacturers into jumping the gun and implement solutions from a future standard never yet agreed to?S l i d e 28Tuesday, 01 February 2000What’s Happening Now?•Document prepared by Working Group 6 using simple limit approach for CENELEC proposed to IEC.•IEC established several new Task forces to address revision.•Latest Proposal now being written as committee Draft.•May be voted on and implemented in time for Jan 1st2001.At last a breakthrough!While I have no comment on the need for the limits or their applicability to the range of products falling within the scope of the standard, if the new proposal goes through we will at least be in abetter position to make realistic tests. S l i d e 29Tuesday, 01 February 2000Main Features of New Draft•Single Limit system for Class C and D–Average value of all measured Harmonic Currents –Maximum Measured value to establish Limits•Removal of Annex C1•Class D Products Defined–600 Watt Maximum / 75 Watt minimum–PC’s, Monitors, TV’s, Video’s, Multimedia*, Printers*,Fax Machines*, *Not Professional equipmentThe new draft offers a significant leap forward in terms of clarification. Not only has it standardised on the single limit approach but many of the other areas of controversy have been eliminated. Most significant is the list of products which fall into Class D. S l i d e 30Tuesday, 01 February 2000Main Features of New Draft•Less Ambiguous–Conflicting Statements removed –Single method using time domain–Annex B extracted from new 61000-4-7•specified power–Manufacturer defines power when near boundaries•i.e.Products near 75 Watt or 600 Watt limits •Allows for production variationsIt is intended that the Manufacturer will have certain discretion in the specifying of power levels etc to avoid confusion when products are close to specification boundary conditions.S l i d e 31Tuesday, 01 February 2000Main Features of New Draft•Better flow graphThe New propsal include a simplified flow graph, completely losing the need to perform the ambiguous waveform shape test.S l i d e 32Tuesday, 01 February 2000Current Equipment•Generally software and Firmware upgrades–Most manufacturers claim equipment is upgradeable•Compatibility of results–For steady state equipments differences are small –Small advantage to products with varying load–Test using measured power for limits where possible andselect single limits if possibleFrom discussion that I have had with several test equipment manufacturers, it is likely that most systems will be modifiable to meet the new requirements.S l i d e 33Tuesday, 01 February 2000Timescales•Committees voting over next two months •Aim to introduce before 1/1/2001•???It’s ironic, if a proper compromise had been found five years ago the standard would be in and working by now! Clearly an indication that the more simple the solution the more likely it is to work.S l i d e 34Tuesday, 01 February 2000Other considerations•Equipment Calibration–Analysers can be calibrated at NPL –Sources still a problem•Need for full Calibration CapabilityCalibration remains an issue for theHarmonic current measuring equipment. In the UK our National Standards Laboratory NPL has developed a calibration service for the analyser part of the equipment. Traceable Calibrations for the sources remain a problem. S l i d e 35Tuesday, 01 February 2000Flicker Standard EN 61000-3-3•Does your product have a peak inrush > 30 Amps?•If yes watch out for some implementations–Already hijacked with introduction of D max limit–Attempts to imply moment of switch on is part of test –Make sure you know how the test is being carried out•Just like with Harmonics testing, due to the poor wording in the standards, accreditation bodies could not ensure a single approach is taken.As an addendum to the Harmonics debacle of the last 5 years the sister standard EN 61000-3-3 is also the subject of some confusion.Biased committee representation was not limited to Harmonics and the same people were responsible for attempting to hijack the Flicker standard and turning it into a Peak Inrush test as well as the existing Flicker requirements.Already different equipment manufacturers have different implementations, some including the moment of equipment switch on in the test period.The effect of this is to cause any product with an inrush current of over about 29 amps to fail the Dmax requirement.Does this affect you?。